Effective management of energy consumption at the consumer level is a vital part of the future smart Grids. An efficient residential energy management system (REMS) is a crucial and necessary. The proposed system also includes a renewable energy source integration module that enables the system to switch between grid and renewable sources based on availability of power source. The REMS integrates ML algorithms with smart home technologies to analyze energy consumption patterns from various power sources and control the renewable power sources utilization and automatically limits heavy loads which are in use during peak hours. Automation of switchover is performed using Artificial Neural Networks (ANN) and Support Vector Machine (SVM), Multiple Linear Regression (MLR) machine learning algorithms for suggesting optimized human-like decisions. The results prove that SVM is superior to ANN in terms of classification accuracy.
1_Introduction + EAM Vocabulary + how to navigate in EAM.pdf
Smart Residential Energy Management System Using Machine Learning.pptx
1. Smart Residential Energy Management Using Machine
Learning
CHAITANYA BHARATHI INSTITUTE OF TECHNOLOGY(A)
DEPARTMENT OF ELECTRICAL & ELECTRONICS ENGINEERING
BE EEE Sem-VII AY 2023-24
Project
Review-2 [6-11-2023]
Project Guide:
Dr. B. Krishna Chaitanya
Presented By:
B. Raju-160120734106
Abdul Javvad Ahmed-160120734083
R.Hareyaank-160120734090
3. Introduction:
• The residential energy management system (REMS) is an advanced
system that has been designed to manage and control the energy
consumption of residential buildings.
• The purpose of the REMS is to optimize energy usage by switching
between grid and renewable energy sources based on demand and
availability.
• The system employs machine learning algorithms to learn from past
usage patterns and to predict future energy requirements.
• This allows the system to make informed decisions about when to
switch between grid and renewable energy sources based on
availability of load.
4. Abstract
• Effective management of energy consumption at the consumer level is a
vital part of the future smart Grids. An efficient residential energy
management system (REMS) is a crucial and necessary. The proposed
system also includes a renewable energy source integration module that
enables the system to switch between grid and renewable sources based
on availability of power source. The REMS integrates ML algorithms with
smart home technologies to analyse energy consumption patterns from
various power sources and control the renewable power sources utilization
and automatically limits heavy loads which are in use during peak hours.
Automation of switchover is performed using Artificial Neural Networks
(ANN) and Support Vector Machine (SVM), Multiple Linear
Regression(MLR) machine learning algorithms for suggesting optimized
human-like decisions. The results prove that SVM is superior to ANN in
terms of classification accuracy.
5. Objectives
• The primary objective of the REMS is to reduce energy costs for
homeowners while minimizing their carbon footprint.
• By using renewable energy sources such as solar panels, wind
turbines, or geothermal heat pumps, the REMS helps to reduce the
dependence on fossil fuels and lower the carbon emissions associated
with energy consumption.
• The REMS is an essential tool for energy management in residential
buildings, as it helps to reduce energy consumption, lower costs, and
promote sustainability.
6. Literature :
S.no Year Author Name Publisher Article Name Outcomes Advantages
1 2017
Krishna
Prakash N,
Prasanna
Vadana D.
IEEE Explore
Machine learning
based Residential
Energy
Management System
An attempt
is made in this paper to
develop an REMS for domestic
installations
that reduces the power consumption
from grid
by effectively switching the loads
between grid and renewably energized
local storage.
REMS when installed in every
home
cumulatively show a significant
reduction
in the burden on the legacy grid
and
improvement in the small scale
penetration
of renewable energy resources
2 2019
Rozeha A.
Rashid,
Leon Chin
IEEE Explore
Machine Learning
for Smart Energy
Monitoring of
Home Appliances
Using IoT
The project's outcomes encompass
improved energy efficiency,
environmental benefits, technological
advancement, user empowerment,
data-driven decision making,
affordability, and education in the
context of home energy management.
1. The project contributes to
the improvement of energy
efficiency in households by
providing users with insights
into their energy
consumption.
7. 3 2020
Mahmood Reaz
Sunny;
Md Ahsan Kabir;
IEEE Explore
Residential Energy
Management: A Machine
Learning Perspective
1.The methodology used in this
article is primarily a literature review
and discussion of key concepts and
research findings related to machine
learning applications in residential
energy management.
2.It involves categorizing machine
learning techniques into supervised,
unsupervised, semi-supervised, and
reinforcement learning.
1.The article informs
readers about the
potential benefits of
applying machine
learning in residential
energy management,
such as improved energy
efficiency and reduced
environmental impact.
2.It highlights the
importance of
collaborative efforts
among consumers,
utilities, industries,
academia, and
government institutions
to overcome challenges
in this domain.
4 2022 Rozeha A. Rashid,
M. Adib Sarijari
IEEE Explore
Machine Learning
Bill Prediction for IoTbased
Utility
Management
System
1.The system uses an ESP32
microcontroller to monitor energy
usage and control appliances.
2.An Artificial Neural Network (ANN)
model, trained with MATLAB,
predicts monthly electricity bills
based on current usage and the day
of the month.
1.Users can make
informed decisions
about energy
conservation based on
bill predictions.
2. The ability to monitor
energy usage in real-
time allows users to
identify energy-intensive
appliances or behaviors
and take immediate
corrective actions.
8. Research Gap/Proposed Technology:
• The ML-based Energy Residential Management System project using
Python can help households to switch between grid and renewable
energy sources based on the load demand and availability of energy
sources.
• Machine learning algorithms can be used to train a model to predict
the power source based on input variables, and how this model can
be used to make decisions in real-time based on the input values.
• By using a combination of data analysis and machine learning
techniques, REMS can help optimize energy usage, reduce costs, and
increase the use of renewable energy sources, thereby promoting
sustainable and efficient energy systems
9. Methodology
• The residential energy management system (REMS) uses machine learning
algorithms to optimize energy usage by switching between grid and renewable
energy sources.
• The rule-based system takes into account several factors, including energy
demand, renewable energy availability, and energy storage capacity.
• When renewable energy sources are available and energy demand is low, the
system switches to using renewable energy sources.
• When renewable energy sources are not available, or when energy demand is
high, the system switches back to using grid energy.
• Based on its charge-discharge transactions and grid availability, a Residential
Energy Management System (REMS) is created to transfer potential loads to
locally stored renewable energy, thereby lowering the total amount of electricity
drawn from the grid.
10. Data Collection:Status Parameters
Parameter Status Description
Grid 0
1
Not Available
Available
Battery SOC -2
-1
0
1
2
0% - 25%
25% - 50%
50% - 75%
75% - 100%
100%
Day/Night 0
1
Not Available
Available
Type of Load 0
1
2
3
Light Loads
Medium Loads
More Medium loads
Heavy Loads
11. Types of Loads
Type of Load Power Consumption
Light Loads 0-150
Medium Loads 150-350
More Medium Loads 350-650
Heavy Loads 650-1000
Decisions Descriptions
1 Maintain same State
2 Connect to Battery
3 Connect to Grid
4 Message to Consumer
Decisions
12. Sample Cases
Case Grid Battery SOC Day/Night Type of Load Decision
a 0 2 1 0 Maintain Same
State
b 1 1 1 1 Connect to
Battery
c 1 1 0 3 Connect to Grid
d 0 0 0 3 Message
Consumer
21. Outcome:
• There are many algorithms in which we can do our project so
rather than using any random algorithm we have gone through
different algorithms.
• We have executed the codes of algorithms and we analyzed there
accuracy.
• As seen from the comparision table above the accuracy of support
vector machine is highest.
• Therefore, for our project we are going to use svm algorithm.
24. SVM
• #importing libraries
• import pandas as pd
• import numpy as np
• from sklearn import datasets
• from sklearn.model_selection import train_test_split
• from sklearn.svm import SVC
• from sklearn.metrics import accuracy_score
• from matplotlib import pyplot as plt
• from mlxtend.plotting import plot_decision_regions
• #data given as csv file for training
• data = pd.read_csv('project.csv')
• #dividing data into dependent and independent variables
• X = data.iloc[:, :-1].values
• y = data.iloc[:, -1].values
• # splitting of data
• X_train, X_test,y_train, y_test = train_test_split(X,
• y, test_size=0.3, random_state=42)
25. • # train svm model
• svm = SVC(kernel='rbf', C=1.0)
• svm.fit(X_train, y_train)
• # predictions using test data
• y_pred = svm.predict(X_test)
• #accuracy
• accuracy = accuracy_score(y_test, y_pred)
• print("Accuracy:", accuracy)
• # Plot decision regions
• plot_decision_regions(X_train.values,y_train.values,
• clf=svm, legend=2)
• plt.xlabel('Grid')
• plt.ylabel('day/night')
• plt.title('SVM Decision Region')
• plt.show()
• #input
• Grid=int(input("Grid availability:"))
• Battery=int(input("Battery State of Charge:"))
28. Conclusion:
• Integrating the REMS system with a smart grid would allow for better
control and optimization of power usage. The system could be designed to
automatically switch between different power sources based on real-time
data from the grid.
• Currently, the system is designed to switch from grid to PV only. In the
future, other renewable energy sources like wind or hydropower could be
integrated into the system to further reduce reliance on non-renewable
sources.
• In conclusion, this research underscores the significance of performing
energy management in smart buildings to achieve sustainability goals,
improve operational efficiency, and enhance overall building performance.
The findings advocate for the widespread adoption of intelligent energy
management systems as a fundamental step towards a greener and more
energy-efficient future.